Prediction Model of Pumpkin Rootstock Seedlings Based on Temperature and Light Responses

Abstract

Temperature and light are the key factors that affect the quality of pumpkin rootstock seedlings’ growth process. Responses to temperature and light are an important basis for optimizing the greenhouse environment. In order to determine the quantitative effects of temperature and light on the growth and development of pumpkin (Cucurbita moschata cv. RTWM6018) rootstock seedlings, relationships between temperature, light, and pumpkin rootstock seedlings growth were established using regression analysis. The results indicated that the daily average temperature had a significant negative correlation with the development time of pumpkin rootstock seedlings, and the shoot dry weight of pumpkin rootstock seedlings increased within a certain range of the daily light integral (DLI). We established a prediction model of pumpkin rootstock seedling quality indicators (hypocotyl length, stem diameter, shoot dry weight, root dry weight, root shoot ratio, and seedling quality index) based on thermal effectiveness and photosynthetic photon flux density (TEP). The coefficient of determinations (R²) of the hypocotyl length and seedling quality index prediction models of pumpkin rootstock seedlings, based on accumulated TEP, were 0.707 and 0.834, respectively. The hypocotyl length and seedling quality index prediction models of pumpkin rootstock seedlings, based on accumulated TEP, were y₁ = 0.001 x² − 0.180 x + 13.057 and y₂ = 0.008 x^0.722, respectively, which could be used for predicting the growth of pumpkin rootstock seedlings grown under different temperature and light conditions.

1. Introduction

In modern agriculture, the vegetable seedling stage and its subsequent cultivation stage are often separated due to planting density, environmental conditions, and management [1]. Prior studies indicated that the seedling quality affects the subsequent growth and mature plants’ yield [2]. Moreover, the demand in China for vegetable seedlings reached 680 billion in 2018 [3]. Thus, cultivating high-quality seedlings is vital for the vegetable seedling industry.

Grafting has been gradually applied in vegetable seedling production to improve seedling quality in terms of plant resistance to pathogens and abiotic stresses [4]. Grafting has long been utilized to improve water and nutrient uptake by strengthening the vascular connection between the scion and rootstock tissues [5]. Suitable rootstocks can be used to absorb mineral nutrients from the soil or substrate, and a previous study [6] indicated that the influences of grafting mainly depend on the rootstock. Pumpkin (Cucurbita moschata Duch.) is one of the most important vegetables and is commercially cultivated worldwide. Asia led in worldwide pumpkin production at 59.9% in 2020, followed by Europe (17.5%) and America (11.9%). In China, the area harvested, and production of pumpkin reached 0.40 million hectares and 7.48 million tones, accounting for 20.0% and 26.8% worldwide, respectively [7]. In addition, pumpkin is often applied as a rootstock in grafted watermelon seedlings [5,8] and grafted cucumber seedlings [9,10].

Crop models, also known as crop growth and development models, can quantitatively describe and predict crop growth and the development process and their relationships with the environment [11]. As for simulation models of the greenhouse crop development period, Perry et al. [12] used the accumulated temperature method to predict the harvest time of cucumber; Uzun et al. [13] established a simple regression model based on mean temperature for predicting the time elapsed from seed sowing to seedling emergence for some vegetable crops (namely, tomato, pepper, aubergine, pea, and carrot); and Sarba et al. [14] established a polynomial mathematical model of the number of days required for tomato seedlings development based on temperature and light intensity. Moreover, artificial intelligence methods (including neural networks, fuzzy logic, and deep learning) [15,16,17] were used to predict lettuce harvest time. Regarding simulation models of greenhouse crop quality index, Hang et al. [18] established a lettuce leaf area model based on the accumulated product of thermal effectiveness and photosynthetically active radiation (TEP) method; Chen et al. [19] proposed a method for obtaining the regulatory target value using the U-chord length curvature method based on the dry-weight model-fitted curve characteristics to obtain the optimum light intensity at the tomato seedling stage; and Liu et al. [20] proposed using a photothermal ratio (PTR) to quantify the effects of PTR during vegetative (PTRv) or reproductive (PTRr) phases on the quality of fully grown ‘Freedom’ poinsettia. These models clarified the relationships between plant growth, development, quality, and environmental conditions, which could be widely used in regulating the temperature and light environment during crop growth. However, prediction models for pumpkin rootstock seedlings’ development time and quality index are rarely reported.

The key indicators affecting grafting are seedling age, hypocotyl length, stem diameter, and seedling quality index [21]. The quality of pumpkin rootstock seedlings significantly influences the subsequent grafting process, such as the grafting survival rate and grafted seedling quality, growth, and yield [5]. Pumpkin rootstock seedlings are affected by various environmental factors during plant growth and development, especially temperature and light. Under high-temperature and low-light conditions, vegetable seedlings exhibit rapid development rates and slender hypocotyls, resulting in decreased dry matter accumulation and seedling quality [2]. In contrast, vegetable seedlings’ relative growth rate and hypocotyl length decrease significantly in winter with low temperature and weak light [22], and the grafting interface may be buried by soil. Until now, quality indices affecting the grafting survival rate and the quality of grafted seedlings, such as the seedling age, hypocotyl length, and stem diameter of pumpkin rootstock seedlings, have not been able to be predicted due to the lack of a development time and quality index prediction model. Thus, quantitative control of the temperature and light environment is of great significance in growing pumpkin rootstock seedlings in a greenhouse.

Considering this, the present study aimed to evaluate the impacts of light intensity and temperature on the growth of pumpkin seedlings and establish models for the environmental factors and representative parameters of pumpkin rootstock seedlings. The results could provide guidelines for regulating the temperature and light environment during the cultivation of pumpkin rootstock seedlings grown in a greenhouse.

2. Materials and Methods

2.1. Plant Materials and Growth Conditions

Pumpkin (Cucurbita maxima cv. RTWM6018) seeds (Golden Ma Ma Agricultural Science & Technology Co., Ltd., Qingdao, China) were sown in 50-cell plug trays (54 cm × 28 cm × 4.5 cm) (Shandong Lige Technology Co., Ltd., Jinan, China) filled with a mixture of vermiculite (Shandong Lige Technology Co., Ltd., Jinan, China), peat (The Pindstrup Group, Kongersle, Denmark), and perlite (Shandong Lige Technology Co., Ltd., Jinan, China) (3:1:1, V/V) in a glass greenhouse with a floor area of 2736 m² at Qingdao Agricultural University (36°19′ N, 120°23′ E), Qingdao, Shandong Province, China. The greenhouse was equipped with a hot-air heating system and a pad-and-fan cooling system to control the temperature between 15 °C and 32 °C. The relative humidity was maintained at 60–70%, and the CO₂ concentration was not controlled during the growth period.

2.2. Collection of Environmental Data

Seventy experiments were conducted in the greenhouse between March 2021 and October 2022, with 3 replicates and 50 seedlings per replicate in each experiment. Environmental data collectors (Hobo, Onset Computer Corp., Bourne, MA, USA) were used to record temperature and light intensity during each experiment (

Table 1. Daily average temperature and DLI during the growth of pumpkin rootstock seedlings in 2021 and 2022.

Years	Variables	Min.	Max.	Average	SD
2021 (n = 48)	Daily average temperature (°C)	16.5	29.3	22.7	3.2
	DLI (mol m−2 d−1)	2.1	16.0	8.6	3.5
2022 (n = 22)	Daily average temperature (°C)	17.8	31.5	23.0	4.2
	DLI (mol m−2 d−1)	3.2	16.6	8.4	3.8

Note: n indicates the number of experiments.

, Figure 1). Eight experiments were randomly selected among the experiments conducted in 2021 to analyze the effects of environmental factors on the growth of pumpkin seedlings. The temperature, light environment, and seedling growth data collected in 2021 were used to establish the prediction model, and the data collected in 2022 were used to verify the prediction model.

2.3. Measurement of Plant Growth Characteristics

The sowing date, seedling emergence date, and seedling date were recorded to calculate pumpkin rootstock seedlings’ emergence time and seedling time in each experiment. Seedling emergence time and the seedling time accounted for the period from sowing to emergence and from emergence to the growth of a seedling with one true leaf, respectively. The hypocotyl length, stem diameter, and dry weight of the pumpkin rootstock seedlings were measured when the seedlings reached the last stage (with one true leaf). The stem diameters of the pumpkin rootstock seedlings were measured with a vernier caliper (DL91150, Deli Tools Co., Ltd., Ningbo, China). The leaves and roots were dried in an oven at 105 °C for 3 h and subsequently left at 80 °C until reaching a constant weight, and then the dry weight of leaves and roots were recorded. The shoot dry weight and root dry weight of the pumpkin rootstock seedlings were measured with an electronic analytical balance. Root shoot ratio and seedling quality index [24] were calculated using the following formula:

$$Seedling quality index=\frac{stem diameter\left(\mathrm{mm}\right)}{plant height\left(\mathrm{cm}\right)}\times total dry weight\left(\mathrm{g}\right)$$

2.4. Establishment of Prediction Model for Pumpkin Rootstock Seedlings

The relationships between the development times (emergence time and seedling time) of pumpkin rootstock seedlings, daily average temperature, and daily light integral were established via stepwise regression of multiple factors and quadratic terms with daily average temperature and daily light integral as variables and emergence time and seedling time of pumpkin rootstock seedlings as dependent variables.

TEP [25,26] is the product of thermal effectiveness (TE) and photosynthetic photon flux density (PPFD). A model for predicting the growth and development of pumpkin rootstock seedlings was established with the accumulated TEP as the independent variable and the seedling quality indicators as the dependent variables. Accumulated TEP was calculated using the daily TEP (DTEP); the calculation formula is as follows:

$$DTE{P}_i={\displaystyle \sum }_{j=1}^n\left(RT{E}_j\times PPF{D}_j\times L\times {10}^{-6}\right)$$

(1)

$$TEP={\displaystyle \sum }_{i=1}^{k}DTE{P}_i$$

(2)

$$RTE=\{\begin{array}{cc}\hfill 0,& T<T_b\hfill \\ \hfill \left(T-T_b\right)/\left(T_{ob}-T_b\right),& T_b\le T<T_{ob}\hfill \\ \hfill 1,& T_{ob}\le T<T_{ou}\hfill \\ \hfill \left(T_m-T\right)/\left(T_m-T_{ou}\right),& T_{ou}\le T\le T_b\hfill \\ \hfill 0,& T\ge T_m\hfill \end{array}$$

(3)

where L is the observation time step and n is the number of observations per day. Data on temperature and light intensity were collected every 10 min in this study; RTE_j and PPFD_j are the relative thermal effect and photosynthetic photon flux density (PPFD; μmol m ⁻² s⁻¹) at time j of any day, respectively; i and k indicate seedling development days and the total days of seedling development, respectively; DTEP is the accumulated the product of the daily thermal effect combined with the photosynthetic photon flux density; T_b, T_ob, T_ou, and T_m are the lower development threshold temperature, the lower optimum temperature threshold, the upper optimum temperature threshold, and the higher development temperature threshold, respectively. The three fundamental temperatures of pumpkin rootstock seedlings are shown in

Table 2. Fundamental temperatures of pumpkin rootstock seedlings.

Temperature Parameters	Fundamental Temperature (°C)
Tb	10
Tob–Tou	25–30
Tm	50

Note: T_b, T_ob, T_ou, and T_m are the lower development threshold temperature, the lower optimum temperature threshold, the upper optimum temperature threshold, and the higher development temperature threshold, respectively.

[27,28,29].

2.5. Model Evaluation

The prediction accuracy of the pumpkin rootstock seedling prediction model was evaluated using the coefficient of determination (R²) of regression analysis and the calculation of root mean square error (RMSE). The calculation formula of RMSE is as follows:

$$RMSE=\sqrt{\frac{1}{n}{\displaystyle \sum }_{i=1}^n{\left(x_{obs,i}-x_{model,i}\right)}^2}$$

(4)

where x_obs,i, x_model,i, and n are the observed value, the predicted value, and the number of observations, respectively.

2.6. Statistical Analysis

One-way ANOVA and stepwise regression analysis of experimental data were carried out using DPS (version 7.05, Hangzhou Ruifeng Information Technology Co., Ltd., Hangzhou, China). Significant differences were considered at p < 0.05. The scatter plot was drawn using Microsoft Excel 2016 software and made to fit the variation curves of the accumulated TEP and pumpkin rootstock seedling quality index. Origin (version 9.8, Origin Lab Corporation, Northampton, MA, USA) was used to compare the relationships between predicted and measured values.

3. Results and Discussion

3.1. Effects of Temperature and Light Intensity on Growth of Pumpkin Rootstock Seedlings

The average daily temperature in the range of 21.7–26.9 °C significantly affected pumpkin rootstock seedlings’ emergence and seedling time (

Table 3. Effects of temperature and light on the morphology of rootstock pumpkin seedlings.

Variables	Daily Average Temperature and Daily Light Integral (°C, mol m−2 d−1)
	T1L1 (21.7,5.6)	T2L2 (24.3,14.5)	T3L3 (24.0,9.7)	T4L4 (23.7,7.2)	T5L5 (25.0,9.3)	T6L6 (22.8,5.2)	T7L7 (25.0,11.2)	T8L8 (26.9,7.8)
Emergence time (d)	4.8 ± 0.3 a	4.1 ± 0.2 b	4.6 ± 0.2 a	4.6 ± 0.2 a	4.1 ± 0.2 b	4.0 ± 0.4 b	3.8 ± 0.3 b	2.8 ± 0.3 c
Seedling time (d)	6.9 ± 0.2 a	5.2 ± 0.3 bc	4.2 ± 0.3 f	5.0 ± 0.4 cd	4.8 ± 0.3 d	5.5 ± 0.4 b	4.7 ± 0.3 de	4.3 ± 0.3 ef
Plant height (cm)	8.2 ± 0.5 b	6.6 ± 0.1 c	6.8 ± 0.1 c	9.7 ± 0.2 a	5.6 ± 0.4 d	8.2 ± 0.8 b	6.0 ± 0.4 cd	9.3 ± 0.7 a
Hypocotyl length (cm)	7.8 ± 0.5 b	6.3 ± 0.2 c	6.5 ± 0.1 c	9.3 ± 0.3 a	4.9 ± 0.5 d	7.9 ± 0.8 b	5.5 ± 0.7 d	8.8 ± 0.7 a
Stem diameter (mm)	4.77 ± 0.14 b	4.27 ± 0.24 cd	5.15 ± 0.31 a	4.25 ± 0.11 cd	4.41 ± 0.32 cd	4.25 ± 0.24 d	4.41 ± 0.27 cd	4.52 ± 0.22 bc
Shoot dry weight (mg)	183.7 ± 17.6 d	220.5 ± 13.9 ab	211.8 ± 16.4 abc	208.0 ± 19.9 bc	191.8 ± 10.7 cd	200.0 ± 13.4 c	233.3 ± 15.3 a	204.2 ± 16.6 bc
Root dry weight (mg)	42.7 ± 2.1 cd	51.0 ± 3.5 b	46.5 ± 1.9 bc	44.6 ± 3.8 c	45.5 ± 4.1 c	33.4 ± 1.9 e	58.4 ± 4.8 a	38.7 ± 2.8 d
Total dry weight (mg)	226.0 ± 18.0 d	269.0 ± 17.6 ab	257.5 ± 17.1 bc	252.6 ± 22.5 bc	235.4 ± 14.9 cd	236.4 ± 15.2 cd	287.9 ± 19.0 a	242.0 ± 17.8 cd
Root shoot ratio	0.199 ± 0.013 bc	0.216 ± 0.016 ab	0.199 ± 0.010 bc	0.215 ± 0.017 ab	0.235 ± 0.016 a	0.187 ± 0.012 c	0.218 ± 0.019 ab	0.190 ± 0.018 c
Seedling quality index	0.143 ± 0.011 cd	0.207 ± 0.016 b	0.189 ± 0.015 b	0.115 ± 0.011 e	0.159 ± 0.009 c	0.117 ± 0.010 e	0.249 ± 0.018 a	0.134 ± 0.006 d

Note: different letters in the same row indicate significant differences based on the least significant difference (LSD) test at p < 0.05.

). The pumpkin rootstock seedlings’ emergence and seedling time decreased gradually with increasing daily average temperature. The pumpkin rootstock seedlings’ emergence and seedling times (2.8 and 4.3 days, respectively) were the shortest under the environment of T8L8.

The pumpkin rootstock seedlings grown under lower and higher DLIs had a shorter plant height and hypocotyl length than seedlings grown under the optimized DLI. The plant height and hypocotyl length of pumpkin rootstock seedlings was the largest under T4L4 and T8L8. The stem diameter of pumpkin rootstock seedlings grown under a higher daily average temperature was smaller than that of seedlings grown under a lower daily average temperature; however, different trends were observed in the stem diameter of pumpkin rootstock seedlings grown under different DLIs. The shoot dry weight, root dry weight, root shoot ratio, total dry weight, and seedling quality index of pumpkin rootstock seedlings grown under T7L7 were higher than those of other pumpkin rootstock seedlings (Table 3).

The development processes of tomato [14], cucumber [30], pepper [31], and lettuce [32] have been reported to be significantly affected by temperature; however, a relationship between temperature and the development time of pumpkin rootstock seedlings has rarely been reported. The hypocotyl of pumpkin rootstock seedlings is an important organ in grafting and is closely related to grafting quality [33]. The hypocotyl length of pumpkin rootstock seedlings grown under DLI of 14.5 mol m⁻² d⁻¹ was lower than that of seedlings grown under a DLI of 5.2 mol m⁻² d⁻¹ (Table 3). Increasing the DLI within a certain range may lead to shorter cell and plant heights, thicker stems, and higher-quality plants. Hwang et al. [34] found the stem length of tomato and red pepper seedlings increased as the DLI increased within a certain range; however, an excessive DLI inhibited the extension of hypocotyl. Kitaya et al. [35] also indicated that the hypocotyl length of lettuce increased as the DLI increased from 5.8 to 8.6 mol m ⁻² d⁻¹. Studies have shown that increasing the DLI is beneficial to increase the dry matter accumulation of seedlings, including lettuce [36], cucumber [37,38,39], and tomato [20] seedlings. The current findings on the development time and quality indicators of pumpkin rootstock seedlings are parallel to those reported by previous researchers.

3.2. Prediction Model of Development Time of Pumpkin Rootstock Seedlings Based on Daily Average Temperature and DLI

The prediction models of pumpkin rootstock seedlings’ emergence and seedling time were y₁ = 0.009 x₁² − 0.720 x₁ + 16.463 and y₂ = 14.781 − 0.405 x₁ + 0.007 x₂², respectively (

Table 4. Regression model of emergence time and seedling time of pumpkin rootstock seedlings based on daily average temperature and daily light integral.

Variables	n	Regression Equation	R2	F Value	p
Emergence Time (d)	48	y1 = 0.009 x12 − 0.720 x1 + 16.463	0.789	242.089	0.0001
Seedling Time (d)	48	y2 = 14.781 − 0.405 x1 + 0.007 x22	0.789	86.025	0.0001

Notes: n indicates the number of experiments; x₁ is the daily average temperature (°C); x₂ is the daily light integral (mol m⁻² d⁻¹).

). The daily average temperature was a key indicator affecting the pumpkin rootstock seedlings’ emergence and seedling time in the regression model, and the daily average temperature and pumpkin rootstock seedlings showed a quadratic relationship. The pumpkin rootstock seedlings’ emergence and seedling time, according to daily average temperature and the DLI, were 78.9% and 78.9%, respectively (Table 4). The daily average temperature and DLI had a high potential for predicting pumpkin rootstock seedlings’ emergence time and seedling time.

Previous researchers [40] used the effective accumulated temperature to establish crop growth and development models to simulate the growth period of crops. The effective temperature accumulation method considered that a certain temperature was required to predict crops’ development stage. This method was applicable to field crops and the environment temperature within the upper and lower limit of the temperature range. However, it produced large errors in predicting the growth period beyond the physiological temperature range. The regression model showed that temperature was the key factor affecting the pumpkin rootstock seedlings’ emergence time. The daily average temperature inversely correlated to the pumpkin rootstock seedlings’ emergence time (Table 4). The pumpkin rootstock seedlings’ emergence and seedling time decreased with the increasing daily average temperature.

Moreover, a nonlinear relationship between the temperature and the emergence time of pumpkin rootstock seedlings was found in our study, which was similar to the emergence time of sugar beet reported by Rimaz et al. [41]. Uzun et al. [14] also established a simple regression model based on the average temperature (D = a − b × T + c × T²) to predict the time from sowing to the emergence of tomato, pepper, eggplant, and other vegetable crops. The regression model of pumpkin rootstock seedling growth time showed that the duration from emergence to the stage with one true leaf decreased linearly with increasing of the daily average temperature (Table 4), which was similar to the results of Sarba et al. [15]. The correlation coefficients of daily average temperature and DLI with the seedling time of pumpkin rootstock seedlings were 0.84 and 0.45 (

Table 5. Path analysis of rootstock pumpkin seedling time on daily average temperature and daily light integral (DLI).

Contribution Factors	Correlation Coefficient	Direct Effect	Through Daily Average Temperature	Through DLI
Daily average temperature	–0.866 **	–1.006		0.147
DLI	–0.45 **	0.219	–0.674

Notes: ** indicates the significance at p ≤ 0.01.

), respectively, which indicates that the daily average temperature and DLI had a significant effect on the pumpkin rootstock seedlings’ seedling time. The direct effect of daily average temperature on the pumpkin rootstock seedlings’ seedling time was greater than that of DLI, as determined through path analysis.

3.3. Development of the Simulation Model for Seedling Quality Indices

The hypocotyl length of pumpkin rootstock seedlings increased with decreasing accumulated TEP, showing a quadratic relationship between the hypocotyl length of pumpkin rootstock seedling and accumulated TEP. However, the stem diameter, shoot dry weight, root dry weight, root shoot ratio, and the seedling quality index of pumpkin rootstock seedlings increased with increasing accumulated TEP, showing a power function relationship for the stem diameter, shoot dry weight, root dry weight, root shoot ratio, and seedling quality index of pumpkin rootstock seedlings with the accumulated TEP (Figure 2). The prediction model of the hypocotyl length of pumpkin rootstock seedlings was y = 0.001 x² – 0.18 x + 13.06, and the coefficient of determination (R²) between accumulated TEP and hypocotyl length of pumpkin rootstock seedlings was 0.707 (Figure 2). However, the coefficient of determinations between accumulated TEP and the diameter of the cotyledon base, shoot dry weight, root dry weight, and root shoot ratio of pumpkin rootstock seedlings was 0.398, 0.383, 0.541, and 0.213, respectively. Our results indicated that the stem diameter, shoot dry weight, root dry weight, and root shoot ratio of pumpkin rootstock seedlings were not suitable for establishing a prediction model based on cumulative TEP according to the lower coefficient of determination. The prediction model of the seedling quality index of pumpkin rootstock seedlings was y = 0.008 x^0.722, and the fitting curve of accumulated TEP and seedling quality index of pumpkin rootstock seedling had a higher R² value (0.834) (Figure 2). Due to the prediction models of the hypocotyl length and the seedling quality index of pumpkin rootstock seedlings having higher R² values, the prediction models of hypocotyl length and seedling quality index of pumpkin rootstock seedlings could be further verified.

TEP was a comprehensive factor of the combinations of temperature, light intensity, and photoperiod. There have been many reports on greenhouse crop models based on TEP [19,42]. Our results suggest that the curve fitting degree between the accumulated TEP and the hypocotyl length of pumpkin rootstock seedlings was higher, but the curve fitting degree between the accumulated TEP and the stem diameter, shoot dry weight, root dry weight, and root–shoot ratio of pumpkin rootstock seedlings was lower. The reason for the lower fitting degree of the prediction model of the stem diameter of pumpkin rootstock seedling was that the stem diameter was smaller at the early stage of seedling. Similarly, the reason for the lower fitting degree of the prediction of shoot dry weight, root dry weight, and root–shoot ratio of pumpkin rootstock seedlings was that there was less biological accumulation in the early stage of seedlings, which affected the accuracy of the seedling prediction model. This result is consistent with the research results of Ming et al. [25] and Zhou et al. [26]. In addition, the curve fitting degree of the prediction model of the seedling quality index of pumpkin rootstock seedlings based on the accumulated TEP was higher, which may be due to the fact that the seedling quality index contained more information on seedlings, especially in terms of dry matter accumulation [43].

3.4. Validation of Prediction Model

The environmental data and seedling quality index of pumpkin rootstock seedlings collected in 2022 were used to verify the prediction model. The emergence time, seedling time, hypocotyl length, and seedling quality index of pumpkin rootstock seedlings were predicted by the prediction model, and the RMSE value was calculated to determine the accuracy of the model. Generally, the closer the predicted value to the observed value and the smaller the RMSE value, the more accurate the prediction ability of the model [44]. The results showed that the RMSE values of seedling emergence time and seedling time were 0.399 and 0.408, respectively. In addition, the RMSE values of hypocotyl length and seedling quality index of pumpkin rootstock seedlings were 0.571 and 0.033, respectively. Small RMSE and high R² values in the validation phase indicate the goodness of fit of the predictive model. The R² values of seedling emergence time, seedling time, hypocotyl length, and seedling quality index of pumpkin rootstock seedlings were 0.936, 0.881, 0.829, and 0.189, respectively (Figure 3).

4. Conclusions

Daily average temperature and the DLI were used as indicators of the development time of pumpkin rootstock seedlings using a regression model, and the accumulated TEP was a potential indicator for predicting the quality index of pumpkin rootstock seedlings. Prediction models of hypocotyl length and seedling quality index based on accumulated TEP for pumpkin rootstock seedlings were established in this study. The prediction model for pumpkin rootstock seedlings’ development time, hypocotyl length, and seedling quality index could be used not only for predicting growth grown under different temperature and light conditions but also to optimize the greenhouse temperature and light environment and to cultivate high-quality seedlings within a certain timeframe to meet grafting needs.